Condensate pump shafting vibration fault diagnosis method and device based on multi-parameter analysis and medium

Abstract

The invention relates to the technical field of water pump fault diagnosis, discloses a condensate pump shafting vibration fault diagnosis method and device based on multi-parameter analysis and a medium, and solves the problems of sensor signal distortion and noise interference of a condensate pump under high-temperature and high-humidity working conditions through multi-source data fusion and environment adaptive compensation. Dynamic time-frequency domain feature extraction and an improved lightweight residual network are adopted, and efficient decoupling and accurate recognition of multi-fault concurrent features are achieved; the fuzzy inference rule base and the shafting fluctuation propagation model are combined, the fault positioning precision and the early fault detection rate are remarkably improved, scientific and reliable technical support is provided for diagnosis and maintenance of the shafting vibration fault of the condensate pump, and safe and stable operation of factory equipment is effectively guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of water pump fault diagnosis, and in particular to a condensate pump shaft system vibration fault diagnosis method, device and medium based on multi-parameter analysis. Background Art

[0002] Among factory equipment, condensate pumps are core equipment for public auxiliary projects. Their shaft systems are in operation under high temperature, high humidity and variable load conditions for a long time. Vibration failures occur frequently and the difficulty of diagnosis is significantly higher than that of conventional pumps. In the existing technology, the diagnostic method for condensate pump shaft vibration mainly relies on single parameter analysis. This includes extracting 1-fold and 2-fold frequency amplitudes to determine imbalance or misalignment, and using a linear combination of bearing temperature and amplitude to set a fixed threshold for alarm. In addition, traditional methods generally use static neural network models or ISO 10816 vibration standards, relying only on single sensor data (such as radial vibration amplitude), and lack adaptability to special operating conditions of condensate pumps (such as start-stop transients and impeller cavitation).

[0003] The existing technology has the following key problems:

[0004] (1) There is a lack of compensation mechanism for sensor signal drift and noise interference in the high temperature and high humidity environment of the condensate pump, resulting in distortion of feature extraction;

[0005] (2) When multiple faults occur simultaneously (such as shaft bending and sealing ring wear), a single frequency domain analysis method cannot decouple the fault characteristics;

[0006] (3) The dynamic correlation between phase difference, torque fluctuation and temperature parameters is not fully utilized, resulting in an early fault missed detection rate of more than 35%. For example, when the condensate pump is running at low load, the traditional method ignores the nonlinear coupling between the torque signal and the vibration phase and mistakenly judges the impeller cavitation as bearing looseness.

[0007] Therefore, a diagnostic method that integrates multi-physical quantity dynamic analysis and environmental adaptive correction is urgently needed to achieve accurate fault identification under complex working conditions. For this purpose, a condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis is proposed. Summary of the invention

[0008] In view of the deficiencies in the prior art, the present invention provides a condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis to solve the background technical problems.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis includes the following steps:

[0011] Step 1: Collect vibration signals through axial vibration sensors and radial vibration sensors, collect torque fluctuation signals through torque meters, collect cooling water temperature through temperature sensors, collect motor current parameters through Hall sensors, and obtain water pump speed data in real time; Step 2: Perform environmental adaptive preprocessing on the vibration signal, including signal correction based on temperature drift compensation and anti-aliasing filtering, to generate a denoised multi-channel time domain signal; Step 3: Adopting the working condition-adaptive time-frequency domain joint analysis method, dynamic wavelet packet decomposition and energy-phase correlation feature extraction are performed on the denoised multi-channel time domain signal to construct a multi-dimensional feature matrix containing energy entropy, phase difference and working condition parameters; Step 4: Input the multi-dimensional feature matrix into the improved lightweight residual network for preliminary fault classification and output the fault probability distribution; Step 5: The torque fluctuation signal collected by the torque meter is integrated through the fuzzy inference engine. The torque fluctuation spectrum characteristics and fault probability distribution obtained after spectrum analysis are combined with the historical fault case library for multi-level decision making to generate the final fault type, confidence level and location information of the fault location; Step 6: Dynamically optimize the alarm threshold based on the equipment's cumulative operating time, load rate, and ambient temperature, and trigger a graded warning strategy.

[0012] Furthermore, the step 1 also includes a sensor self-test process:

[0013] Through the coherence function of vibration signal and torque fluctuation signal The amplitude at the power frequency determines the health status of the sensor. , the sensor is marked as abnormal and redundant channel data is enabled. The value range of A is .

[0014] Furthermore, the specific implementation method of the temperature drift compensation in step 2 is:

[0015] According to the mapping relationship between the cooling water temperature and the thermal expansion coefficient of the shafting material, the temperature compensation coefficient of the vibration signal is calculated. , the expression is:

[0016]

[0017] Where γ is the thermal expansion coefficient of the shafting material, T current is the cooling water temperature, is the median temperature calculated by the sliding window.

[0018] Furthermore, the operations of performing dynamic wavelet packet decomposition and energy-phase correlation feature extraction in step 3 include:

[0019] According to the real-time speed change rate δ, δ= , dynamically select the wavelet basis function: when When B is in the range of 3% to 5%, Daubechies 6 wavelet basis is selected; otherwise, Symlet 4 wavelet basis is selected; is the speed difference, is the speed detection cycle time; the decomposition layer number is adaptively adjusted according to the speed frequency band proportion, and the calculation formula of the layer number L is:

[0020]

[0021] in, is the sampling frequency, It is the maximum analysis frequency corresponding to the current speed.

[0022] Furthermore, the improvement of the lightweight residual network includes:

[0023] The dual-path attention mechanism is embedded in the residual block to perform feature reweighting on the channel dimension and spatial dimension respectively. The calculation formula is:

[0024]

[0025] Among them, GAP is the global average pooling, is the channel dimension weight matrix of the fully connected layer, W 2 is the spatial dimension weight matrix of the fully connected layer, ReLU is the ReLU function, is the standard deviation of the real-time feature;

[0026] The pre-trained large ResNet-50 model is compressed into a lightweight network using the knowledge distillation method.

[0027] Furthermore, the fuzzy inference engine in step 5 includes a condensate pump-specific rule base:

[0028] Rule 1: If the high-frequency energy entropy increases suddenly > H 1 And the torque fluctuation 3 times frequency amplitude ratio> H 2 , it is judged to be an impeller cavitation fault, and the confidence level is H 3 , H 1 The value range is 35% to 40%, H 2 The value range is 20% to 25%, H 3 The value range is 0.82~0.85;

[0029] Rule 2: When the axial / radial vibration energy ratio > N 1 And the phase difference fluctuation variance> N 2 When the bearing wear detection submodule is triggered, N1 The value range of N is 2.25~2.35, 2 The value range is 9.5°²~11°²;

[0030] Rule 3: If the cooling water temperature rise rate>S 1 Accompanying current harmonic distortion rate> S 2 , it is judged that the mechanical seal fails, S 1 The value range is 1.3℃ / min~1.6℃ / min, S 2 The value range is 7.5% to 8.5%.

[0031] Furthermore, the generation of the location information of the fault location in step 5 is realized by using the shaft system wave propagation model: according to the peak time difference of the cross-correlation function of the vibration signals at both ends of the shaft , calculate the distance from the fault point to the end of the shaft system connecting the condensate pump and the motor :

[0032]

[0033] Where v is the stress wave velocity in the shaft material and δ is the sensor installation spacing correction.

[0034] Furthermore, the method for dynamically optimizing the alarm threshold in step 6 includes:

[0035] Calculate threshold baseline based on Weibull distribution model , whose expression is:

[0036]

[0037] Among them, α is the shape parameter of equipment life distribution, β is the scale parameter of equipment life distribution, η is the working condition correction factor, t is time, and e is a natural constant;

[0038] Combining the standard deviation σ of real-time features with the threshold baseline Generate final alarm threshold :

[0039]

[0040] Where k is the safety factor.

[0041] The present invention also provides a condensate pump shaft system vibration fault diagnosis device based on multi-parameter analysis, comprising one or more processors for implementing the condensate pump shaft system vibration fault diagnosis method based on multi-parameter analysis as described above.

[0042] The present invention also provides a readable storage medium having a program stored thereon, and when the program is executed by a processor, the above-mentioned condensate pump shaft system vibration fault diagnosis method based on multi-parameter analysis is implemented.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention solves the problems of sensor signal distortion and noise interference of condensate pumps under high temperature and high humidity conditions through multi-source data fusion (vibration, torque, temperature, current) and environmental adaptive compensation (temperature drift correction, anti-aliasing filtering); adopts dynamic time-frequency domain feature extraction (wavelet packet decomposition, energy-phase correlation analysis) and improved lightweight residual network to achieve efficient decoupling and accurate identification of concurrent features of multiple faults; combines fuzzy reasoning rule library with shaft system wave propagation model to significantly improve fault location accuracy and early fault detection rate, providing scientific and reliable technical support for the diagnosis and maintenance of condensate pump shaft system vibration faults, and effectively ensuring the safe and stable operation of power plant equipment.

[0045] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The present invention is a flow chart of the condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis.

[0047] Figure 2 It is a structural schematic diagram of a condensate pump shaft system vibration fault diagnosis device based on multi-parameter analysis of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this technical field without creative work are within the scope of protection of the present invention.

[0049] See also Figure 1 A condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis includes the following steps:

[0050] Step 1: Collect vibration signals through axial vibration sensors and radial vibration sensors, collect torque fluctuation signals through torque meters, collect cooling water temperature through temperature sensors, collect motor current parameters through Hall sensors, and obtain water pump speed data in real time.

[0051] Perform a self-test on the sensor. The process includes:

[0052] Through the coherence function of vibration signal and torque fluctuation signal The amplitude at the power frequency determines the health status of the sensor. , the sensor is marked as abnormal and redundant channel data is enabled. The value range of A is , preferably 0.6.

[0053] Step 2: Perform environmental adaptive preprocessing on the vibration signal, including signal correction based on temperature drift compensation and anti-aliasing filtering, to generate a denoised multi-channel time domain signal. Anti-aliasing filtering is a well-known technology and will not be described in detail.

[0054] Among them, the specific implementation method of temperature drift compensation is:

[0055] According to the mapping relationship between the cooling water temperature and the thermal expansion coefficient of the shafting material, the temperature compensation coefficient of the vibration signal is calculated. , the expression is:

[0056]

[0057] Where γ is the thermal expansion coefficient of the shafting material, T current Cooling water temperature is the median temperature calculated by the sliding window.

[0058] Step 3: Adopt the working condition adaptive time-frequency domain joint analysis method to perform dynamic wavelet packet decomposition and energy-phase correlation feature extraction on the denoised multi-channel time domain signal, and construct a multi-dimensional feature matrix including but not limited to energy entropy, phase difference and working condition parameters.

[0059] The operations of dynamic wavelet packet decomposition and energy-phase correlation feature extraction include:

[0060] According to the real-time speed change rate δ, δ= , dynamically select the wavelet basis function: when When B is in the range of 3% to 5%, preferably 5%, Daubechies 6 wavelet basis is used; otherwise, Symlet 4 wavelet basis is used; is the speed difference, is the speed detection cycle time. The number of decomposition layers is adaptively adjusted according to the speed band ratio, and the calculation formula of the number of layers L is:

[0061]

[0062] in, is the sampling frequency, It is the maximum analysis frequency corresponding to the current speed.

[0063] Step 4: Input the multi-dimensional feature matrix into the improved lightweight residual network (Light-ResNet) for preliminary fault classification and output the fault probability distribution. The working process of the lightweight residual network is a well-known technology and will not be described in detail.

[0064] Among them, the improvements of lightweight residual network include:

[0065] The dual-path attention mechanism is embedded in the residual block to perform feature reweighting on the channel dimension and spatial dimension respectively. The calculation formula is:

[0066]

[0067] Among them, GAP is the global average pooling, is the channel dimension weight matrix of the fully connected layer, W 2 is the spatial dimension weight matrix of the fully connected layer, ReLU is the ReLU function, is the standard deviation of the real-time feature.

[0068] The trained large ResNet-50 model is compressed into a lightweight network using the knowledge distillation method.

[0069] Step 5: The torque fluctuation signal collected by the torque meter is integrated through the fuzzy inference engine. The torque fluctuation spectrum characteristics and fault probability distribution obtained after spectrum analysis are combined with the historical fault case library for multi-level decision-making to generate the final fault type, confidence and location information.

[0070] Among them, the fuzzy reasoning engine contains a special rule base for condensate pumps:

[0071] Rule 1: If the high-frequency energy entropy increases suddenly > H 1 And the torque fluctuation 3 times frequency amplitude ratio> H 2 , it is judged to be an impeller cavitation fault, and the confidence level is H 3 , H 1 The value range is 35% to 40%, preferably 40%. 2 The value range is 20% to 25%, preferably 25%, H 3 The value range is 0.82 to 0.85, preferably 0.85;

[0072] Rule 2: When the axial / radial vibration energy ratio > N 1And the phase difference fluctuation variance> N 2 When the bearing wear detection submodule is triggered, N 1 The value range of N is 2.25 to 2.35, preferably 2.3. 2 The value range is 9.5°²~11°², preferably 10°²;

[0073] Rule 3: If the cooling water temperature rise rate>S 1 Accompanying current harmonic distortion rate> S 2 , it is judged that the mechanical seal fails, S 1 The value range is 1.3℃ / min~1.6℃ / min, preferably 1.5℃ / min, S 2 The value range is 7.5% to 8.5%, preferably 8%.

[0074] Among them, the location information of the fault location is realized by using the shaft wave propagation model: according to the peak time difference of the cross-correlation function of the vibration signals at both ends of the shaft Calculate the distance from the fault point to the end of the shaft system connecting the condensate pump and the motor :

[0075]

[0076] in, is the peak time difference of the cross-correlation function of the vibration signals at both ends of the shaft, v is the stress wave velocity in the shaft material, and δ is the sensor installation spacing correction.

[0077] Step 6: Dynamically optimize the alarm threshold based on the accumulated running time of the equipment, load rate and ambient temperature, and trigger the graded warning strategy. The graded warning strategy is to issue warnings based on different alarm thresholds. The graded warning strategy is a well-known technology and will not be elaborated on.

[0078] The method of dynamically optimizing the alarm threshold includes:

[0079] Calculate threshold baseline based on Weibull distribution model , whose expression is:

[0080]

[0081] Among them, α is the shape parameter of the equipment life distribution, β is the scale parameter of the equipment life distribution, η is the operating condition correction factor, t is time, and e is a natural constant.

[0082] Combining the standard deviation σ of real-time features with the threshold baseline Generate final alarm threshold :

[0083]

[0084] Among them, k is the safety factor, and its value range is 2.0-3.0.

[0085] See also Figure 2 An embodiment of the present invention provides a condensate pump shaft system vibration fault diagnosis device based on multi-parameter analysis, including one or more processors for implementing a condensate pump shaft system vibration fault diagnosis method based on multi-parameter analysis in the above embodiment.

[0086] An embodiment of the condensate pump shaft vibration fault diagnosis device based on multi-parameter analysis of the present invention can be applied to any device with data processing capabilities, and the device with data processing capabilities can be a device or apparatus such as a computer. The device embodiment can be implemented through software, or through hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capabilities in which it is located reading the corresponding computer program instructions in the non-volatile memory into the internal memory for execution. From a hardware perspective, if Figure 2 As shown in the figure, it is a hardware structure diagram of any device with data processing capability in which a condensate pump shaft vibration fault diagnosis device based on multi-parameter analysis of the present invention is located. Figure 2 In addition to the processor, memory, network interface, and non-volatile memory shown, any device with data processing capabilities in the embodiments may also include other hardware, usually based on the actual functions of the device with data processing capabilities, which will not be described in detail.

[0087] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0088] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0089] An embodiment of the present invention further provides a readable storage medium having a program stored thereon. When the program is executed by a processor, a condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis in the above embodiment is implemented.

[0090] The readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The readable storage medium may also be an external storage device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. equipped on the device. Furthermore, the readable storage medium may also include both an internal storage unit of any device with data processing capability and an external storage device. The readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0091] Example

[0092] See also Figure 1 A condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis in the present invention includes the following contents:

[0093] 1. System hardware configuration and data acquisition

[0094] The implementation object of this embodiment is a vertical condensate pump (model: 9LDTN-6P) supporting a 1000MW unit in a power plant, whose shaft system structure includes a four-stage impeller, an intermediate bearing seat and a hydraulic coupler. The hardware system includes a multi-source sensor array and a data synchronization device. The multi-source sensor array includes a vibration monitoring unit, a torque monitoring unit, a temperature monitoring unit and an electrical parameter unit.

[0095] Vibration monitoring unit: Three-way vibration sensors (PCB 352C33, measuring range ±50g, sampling frequency 10kHz) are installed at the driving end and non-driving end of the pump shaft, with an axial spacing of 1.2m and a radial installation angle interval of 120°.

[0096] Torque monitoring unit: A strain gauge torque meter (HBM T40B, accuracy ±0.1%FS) is installed on the output shaft of the hydraulic coupling to collect torque fluctuation signals in real time.

[0097] Temperature monitoring unit: PT100 platinum resistor (accuracy ±0.5℃) is arranged in the bearing seat and cooling water circuit, with a sampling interval of 1 second.

[0098] Electrical parameter unit: The three-phase current of the motor is collected through the Hall sensor (LEM ITC 200-S), and the output frequency of the inverter is obtained synchronously.

[0099] Data synchronization device: PXIe-8840 controller equipped with synchronous acquisition card (NI 9234) is used to achieve strict time alignment of multi-channel signals with synchronization error <1μs.

[0100] 2. Implementation details of environment adaptive preprocessing

[0101] Step 1: Temperature Drift Compensation

[0102] Establishment of thermal expansion coefficient database of shafting material (martensitic stainless steel 2Cr13): measured by laboratory temperature control box .

[0103] Reference temperature Dynamic update: Calculate the sliding median value based on the cooling water temperature data for 12 hours before the current moment. For example, when the historical temperature sequence is [45.3℃, 46.1℃, 47.5℃, 45.8℃], take the median value of 46.1℃.

[0104] Signal correction example: If the current temperature , then the compensation coefficient , scale the amplitude of the original vibration signal.

[0105] Step 2: Anti-aliasing filter design

[0106] According to the pump shaft gear teeth number k=28 and the maximum speed , calculate the cutoff frequency :

[0107]

[0108] An 8th-order Butterworth filter is used with a stopband attenuation of >60dB to ensure that the signal is alias-free in the range of 0 to 11.2kHz.

[0109] Step 3: Sensor self-check and redundancy switching

[0110] Calculate the vibration signal (channel CH1) and torque signal (channel CH2) at the power frequency The coherence function at :

[0111]

[0112] in, is the autopower spectrum, is the cross power spectrum.

[0113] like , determine that CH1 is abnormal, and automatically switch to the backup vibration sensor CH3.

[0114] 3. Dynamic time-frequency domain feature extraction process

[0115] Step 1: Condition-adaptive wavelet packet decomposition

[0116] When the speed change rate Rated speed / second (rated speed 2975r / min corresponds to 5% of 148.75r / min / s), select Daubechies 6 wavelet basis.

[0117] Decomposition layer calculation: current speed corresponds to the maximum analysis frequency Sampling frequency ,but:

[0118]

[0119] 64 sub-bands are generated, covering the range 0-156.25 Hz.

[0120] Step 2: Energy-phase correlation analysis

[0121] Calculate the energy entropy of the k=24 sub-band (corresponding to 75-78.125Hz) :

[0122]

[0123] Extract the phase difference sequence of axial and radial vibration signals , calculate its variance .

[0124] 4. Light-ResNet Fault Classification Implementation

[0125] Network structure parameters:

[0126] Input layer: 64×64 feature matrix (corresponding to 64 sub-bands × 64 time frames);

[0127] Residual block configuration: 4 dual-path attention residual blocks, channel number [64, 128, 256, 512];

[0128] Attention mechanism: global average pooling followed by two layers of full connection (512→32→512), compression ratio 16:1, LeakyReLU negative slope 0.2;

[0129] Knowledge distillation: The teacher network (ResNet-50) is pre-trained on the CWRU bearing dataset, and the student network retains the feature mapping relationship of the first 3 layers of the teacher network.

[0130] Training process:

[0131] Dataset: Three years of operation data of condensate pumps were collected, including 12 types of failure modes (such as impeller cavitation, bearing spalling, shaft bending, etc.), with a total of 8,500 sets of samples;

[0132] Optimizer: AdamW (learning rate 3e-4, weight decay 0.01);

[0133] Output result: The current sample fault probability distribution is [impeller cavitation: 0.72, bearing wear: 0.18, normal: 0.10].

[0134] 5. Fuzzy reasoning multi-level decision-making example

[0135] Rule triggering example:

[0136] High-frequency energy entropy sudden increase detection: current E_{24}=2.37, baseline E_{24_base}=1.68, sudden increase (2.37-1.68) / 1.68≈41% >40%.

[0137] Torque 3-fold frequency analysis: Perform FFT on the torque signal and calculate the amplitude ratio of the 3-fold frequency (149.4Hz) = 28.5% > 25%.

[0138] Judgment result: Trigger rule 1 and output “impeller cavitation fault” (confidence 0.85×0.72=0.612).

[0139] Fault location calculation:

[0140] The time difference between the sensors at both ends of the axis receiving the shock wave , the sound velocity of the shaft material v=5100m / s, and the correction amount δ=0.15m.

[0141]

[0142] The fault point is located 1.84m away from the driving end, corresponding to the position of the second-stage impeller.

[0143] 6. Dynamic threshold optimization and early warning

[0144] Threshold calculation example:

[0145] The equipment has been running for a total of t=25,000 hours, the Weibull parameter α=35,000 hours, β=2.3, and the operating factor η=0.92.

[0146]

[0147] The current characteristic standard deviation σ=0.18, safety factor k=2.5:

[0148]

[0149] The measured characteristic value is 1.23> , triggering a second-level warning (yellow alarm), prompting maintenance to be arranged within 72 hours.

[0150] 7. Implementation effect verification

[0151] The system of this embodiment is deployed on 6 condensate pumps of the same type in a power plant, and compared with the traditional method:

[0152] Missed detection rate: reduced from 35.2% to 6.8% (through continuous monitoring of 12 early cavitation failures over 3 months);

[0153] False positive rate: reduced from 28.7% to 9.4% (eliminated 11 false alarms caused by temperature drift);

[0154] Positioning accuracy: The average error is improved from ±0.8m to ±0.2m (verified by actual measurement using a laser plummet).

[0155] This embodiment achieves accurate diagnosis and positioning of condensate pump shaft vibration faults through multi-source data fusion (vibration, torque, temperature, current), environmental adaptive compensation (temperature drift correction, anti-aliasing filtering), dynamic time-frequency domain feature extraction (wavelet packet decomposition, energy-phase correlation analysis), and improved lightweight residual network and fuzzy reasoning collaborative decision-making, significantly reducing the missed detection rate (from 35.2% to 6.8%) and misjudgment rate (from 28.7% to 9.4%), and improving the fault location accuracy to ±0.2 meters through the shaft wave propagation model, providing scientific and reliable technical support for power plant equipment maintenance.

[0156] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis, characterized in that: The following steps are involved: Step 1: Collect vibration signals through axial vibration sensors and radial vibration sensors, collect torque fluctuation signals through torque meters, collect cooling water temperature through temperature sensors, collect motor current parameters through Hall sensors, and obtain water pump speed data in real time; Step 2: Perform environmental adaptive preprocessing on the vibration signal, including signal correction based on temperature drift compensation and anti-aliasing filtering, to generate a denoised multi-channel time domain signal; Step 3: Adopting the working condition-adaptive time-frequency domain joint analysis method, dynamic wavelet packet decomposition and energy-phase correlation feature extraction are performed on the denoised multi-channel time domain signal to construct a multi-dimensional feature matrix containing energy entropy, phase difference and working condition parameters; Step 4: Input the multi-dimensional feature matrix into the improved lightweight residual network for preliminary fault classification and output the fault probability distribution; Step 5: The torque fluctuation signal collected by the torque meter is integrated through the fuzzy inference engine. The torque fluctuation spectrum characteristics and fault probability distribution obtained after spectrum analysis are combined with the historical fault case library for multi-level decision making to generate the final fault type, confidence level and location information of the fault location; Step 6: Dynamically optimize the alarm threshold based on the equipment's cumulative operating time, load rate, and ambient temperature, and trigger a graded warning strategy.

2. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The step 1 also includes the sensor self-test process: Through the coherence function of vibration signal and torque fluctuation signal The amplitude at the power frequency determines the health status of the sensor. , the sensor is marked as abnormal and redundant channel data is enabled. The value range of A is .

3. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The specific implementation method of the temperature drift compensation in step 2 is: According to the mapping relationship between the cooling water temperature and the thermal expansion coefficient of the shafting material, the temperature compensation coefficient of the vibration signal is calculated. , the expression is: Where γ is the thermal expansion coefficient of the shafting material, T current is the cooling water temperature, is the median temperature calculated by the sliding window.

4. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The operations of dynamic wavelet packet decomposition and energy-phase correlation feature extraction in step 3 include: According to the real-time speed change rate δ, δ= , dynamically select the wavelet basis function: when When B is in the range of 3% to 5%, Daubechies 6 wavelet basis is selected; otherwise, Symlet 4 wavelet basis is selected; is the speed difference, is the speed detection cycle time; the decomposition layer number is adaptively adjusted according to the speed frequency band proportion, and the calculation formula of the layer number L is: in, is the sampling frequency, It is the maximum analysis frequency corresponding to the current speed.

5. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The improvements of the lightweight residual network include: The dual-path attention mechanism is embedded in the residual block to perform feature reweighting on the channel dimension and spatial dimension respectively. The calculation formula is: Among them, GAP is the global average pooling, is the channel dimension weight matrix of the fully connected layer, W2 is the spatial dimension weight matrix of the fully connected layer, ReLU is the ReLU function, is the standard deviation of the real-time feature; The pre-trained large ResNet-50 model is compressed into a lightweight network using the knowledge distillation method.

6. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The fuzzy inference engine in step 5 includes a condensate pump-specific rule base: Rule 1: If the sudden increase in high-frequency energy entropy is greater than H1 and the torque fluctuation 3-fold frequency amplitude ratio is greater than H2, it is determined to be an impeller cavitation fault with a confidence level of H3. The value range of H1 is 35% to 40%, the value range of H2 is 20% to 25%, and the value range of H3 is 0.82 to 0.

85. Rule 2: When the axial / radial vibration energy ratio > N1 and the phase difference fluctuation variance > N2, the bearing wear detection submodule is triggered. The value range of N1 is 2.25 to 2.35, and the value range of N2 is 9.5°² to 11°². Rule 3: If the cooling water temperature rise rate > S1 and the current harmonic distortion rate > S2, it is judged as a mechanical seal failure. The S1 value range is 1.3℃ / min~1.6℃ / min, and the S2 value range is 7.5%~8.5%.

7. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The generation of the location information of the fault location in step 5 is realized by using the shaft system wave propagation model: according to the peak time difference of the cross-correlation function of the vibration signals at both ends of the shaft , calculate the distance from the fault point to the end of the shaft system connecting the condensate pump and the motor : Where v is the stress wave velocity in the shaft material and δ is the sensor installation spacing correction.

8. The condensate pump shaft vibration fault diagnosis method based on multi-parameter analysis according to claim 1 is characterized in that: The method for dynamically optimizing the alarm threshold in step 6 includes: Calculate threshold baseline based on Weibull distribution model , whose expression is: Among them, α is the shape parameter of equipment life distribution, β is the scale parameter of equipment life distribution, η is the working condition correction factor, t is time, and e is a natural constant; Combining the standard deviation σ of real-time features with the threshold baseline Generate final alarm threshold : Where k is the safety factor.

9. A condensate pump shaft vibration fault diagnosis device based on multi-parameter analysis, characterized in that: It comprises one or more processors for implementing a condensate pump shaft system vibration fault diagnosis method based on multi-parameter analysis as described in any one of claims 1-8.

10. A readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a condensate pump shaft system vibration fault diagnosis method based on multi-parameter analysis as described in any one of claims 1-8 is implemented.

Images